Abstract
Background:
Vascular access is a critical component of emergency department (ED) care. Ultrasound guided placement of peripheral intravenous (USIV) catheters is increasingly common. However, USIV are thought to suffer from reduced durability and higher complication rates. Extended dwell catheters (EDC) are long peripheral IVs placed under combined ultrasound and wire guidance. The goal of this study is to compare dwell times and complication rates of EDC to standard peripheral USIV.
Methods:
We performed a retrospective cohort study at a tertiary care adult ED comparing IV placements during a 17-month period (8/1/2018 – 12/31/2019), stratified by standard USIV versus EDC. The primary outcome was catheter dwell time and secondary outcomes included need for inpatient vascular access team (VAST) consultation, peripherally inserted central catheter (PICC) insertions, and radiocontrast extravasations. Multivariable Cox regression time-to-event analyses were used to evaluate dwell times, adjusting for age, gender, BMI and end-stage renal disease.
Results:
359 EDC and 4190 standard USIV were included for analysis. Most USIV (95.6%) and EDC (98.3%) were placed by ED technicians trained in ultrasound vascular access. EDC median dwell time (5.9 days [95%CI: 5.1–6.7]) exceeded standard USIV (3.8 days [95% CI: 3.6–4.0]). Patients with EDC placed in the ED required less VAST consultation (0.84 vs 0.99 charges/encounter), had similar rates of PICC line use (8.0% vs 8.4% of encounters) and had no radiocontrast extravasation events. Multivariable Cox regression demonstrated survival benefit (longer dwell time) favoring EDC (HR 0.70 [95%CI 0.60–0.81]).
Conclusion:
Use of EDC results in longer dwell time and reduces subsequent use of vascular access resources, while maintaining low complication rates. EDC demonstrate superior durability which may justify their selection over standard USIV in some patients.
Keywords: IV, intravenous, peripheral, ultrasound, guided, extended, dwell, catheter
1. Introduction
1.1. Background
When intravenous (IV) access is required, difficult access complicates approximately 10% of all emergency department (ED) visits and can lead to delays in urgent diagnostic testing or treatment1. Ultrasound guided IV (USIV) placement is increasingly common and for many patients can reduce the number of insertion attempts or the need for central venous access2. However, USIV have also been found to have higher rates of premature failure and thus significantly lower dwell times when compared to standard IVs placed with physical landmarks3,4. USIV are also thought to be associated with higher rates of complications including extravasation of computed tomography (CT) iodinated contrast agents5,6. This has led to blanket policies at some institutions prohibiting the use of USIVs for power injection and administration of potentially irritant or vesicant medicines7. Extended dwell catheters (EDC) are long length, ultrasound guided peripheral catheter systems with the addition of wire guidance that are approved for dwell times of up to 29 days. Use of a combined ultrasound and wire guidance system, similar to other Seldinger technique systems for vascular access, is thought to enhance catheter safety via reduction in venous trauma during placement. Additionally, the longer length of EDC (typically 6–10 cm) compared to standard long USIV (4–6 cm) contributes to their durability8. EDC are also different from midline catheters as they are typically inserted below the antecubital fossa, generally have a shorter catheter length, and their final placement does not approach the central veins9. EDC use however, has been limited by their cost relative to standard USIV as well as additional training and time required for their insertion. Lack of evidence as to their durability and reliability has also limited their use.
1.2. Objectives
Our medical center implemented an ED difficult vascular access team led by paramedic trained technicians beginning in 2008 and began utilizing EDCs in 2018. The purpose of this study is to evaluate the comparative durability of EDC to standard USIV, by assessing differences in dwell times and complication rates between catheters.
2. Methods
2.1. Study Design
This study is a retrospective cohort study utilizing electronic medical record data to evaluate the durability and complication rates of peripheral IVs placed in the emergency department with ultrasound guidance. The study timeframe spanned 17-months (08/01/2018 – 12/31/2019). In the two months prior to this period, the Endurance Extended Dwell Peripheral Catheter System (Figure S1, Teleflex, Morrisvile, NC) was introduced in our department in 20 and 22 gauge diameters as well as 6 and 8 cm lengths for both sizes. These EDC were inserted using standard short axis ultrasound guidance with a wire guide system. During the study period, standard USIV were also inserted with similar ultrasound guidance technique. These catheters were BD Insyte Autoguard IV catheters (Becton, Dickinson and Co., Franklin Lakes, NJ) available in 14 to 24 gauge sizes and lengths from 2 to 4.5 cm. Both catheters are standard polyurethane plastic IV catheters.
2.2. Setting, Participants and IV Insertion
The study was conducted at an urban, university affiliated, teaching hospital with 106,418 adult ED visits and 41,081 patients admitted from the ED during the study period (Figure 1). Our institutional IRB approved this study. General peripheral IV access insertion may be performed by nursing staff, paramedic trained ED technicians or physicians and physician assistants. The ED difficult IV access (DIVA) team, consisting of a subset of ED technicians, is available at all hours and by department policy may be called to place ultrasound guided IVs in cases of multiple prior attempts by the bedside nurse or known difficult IV access. All DIVA team inserters were highly experienced with more than 1,000 peripheral IV placements each. The DIVA team training program began in 2008 and well established prior to the study timeframe. Physician and nurse IV inserters had variable levels of experience but consisted of a tiny fraction of operators placing IVs using ultrasound guidance in this study. All IVs placed in patients with ultrasound guidance and admitted from the ED during the study period were selected for inclusion and none were excluded from the initial analysis. Selection of EDC or standard USIV was at the discretion of the inserter but EDC selection generally favors patients with suspected very difficult IV access. However, both groups are considered to have difficult IV access. Departmental policy for DIVA team members mandated insertion of EDC in the forearm. Other inserters (primarily physicians) could place EDC at other sites. Standard USIV site selection was at the discretion of the inserter. For all inserters, departmental policy is for insertion to a depth no more than one third of the catheter length.
Figure 1.
Study flow diagram. All ultrasound guided intravenous catheters (USIV) placed between 8/1/2018 and 12/31/2019 were utilized in the primary analysis. Placement of either standard USIV or extended dwell catheters (EDC) was considered difficult intravenous access.
2.3. Data Collection and Variables
Our institution utilizes the Epic electronic medical record (EMR, Epic Systems, Verona, WI) and standard lines, drains and airways flowsheet charting system for all peripheral IVs placed in the hospital and this data is routinely collected during the course of care. The Epic Clarity data warehouse is also utilized to store clinical information as an Oracle database allowing direct queries for variables entered into the EMR as structured data. Variables from initial placement including time, performing provider, type of IV, length, size, guidance used, number of insertion attempts, and site were extracted directly from Epic Clarity. Removal variables including time, removal reason, and complications were also extracted. Both insertion and removal variables are maintained as a quality analytics report at our institution. Using unique encounter and patient identifiers, these data were joined to variables regarding the ED visit and hospitalization as well as patient characteristics which were also directly extracted from our institution’s research data warehouse. Manual chart review was not performed and data extraction, cleaning and management was performed by the first author who was not blinded to the hypothesis of this study. These data included records of inpatient vascular access services team (VAST) charges, peripherally inserted central catheter (PICC) line insertion, computed tomography (CT) contrast injections and extravasations. All data cleaning and exploratory analyses were performed in Tableau or Tableau Prep Builder (Tableau Software, Seattle, WA). The primary exposure variable in this study was the type of peripheral IV placed (USIV or EDC) and each peripheral IV was considered a single observation.
2.4. Determination of Peripheral IV Failure
The primary outcome of this study was time to IV catheter failure and was defined as removal prior to 8 hours before discharge from the hospital. In cases where no specific removal time was documented, the IV was determined to have remained in place without failure till discharge if charting in the IV maintenance or infusions flowsheets remained. In cases where neither removal time nor maintenance or infusion documentation occurred during the hospitalization, the IV was considered to have failed early; similar to those with dwell time of less than 4 hours.
2.5. Study Size
No a priori sample size estimate was performed as the study time frame was determined by convenience. However, when evaluated post hoc based on a relative 9:1 ratio of standard USIV to EDC placement during the two-month trial period, a desired beta of 0.2 and alpha of 0.05, we estimated that 1751 IV failure events were required for adequate power when dwell time was analyzed by survival in a Cox proportional hazards model with a hazard ratio of 0.8 favoring EDC.
2.6. Statistical Analysis
Descriptive statistics were calculated on all variables of interest. Counts with percentages are included for all categorical variables and medians with interquartile ranges for all continuous variables. Differences in patient characteristics that were continuous variables were assessed for statistical significance using two sample t-tests and for categorical variables the chi-squared or Fisher’s exact test were used where appropriate. Kaplan-Meier survival analysis was used to visualize survival curves and estimate time to IV failure. Survival curves were evaluated for statistically significant differences using the log rank test. Multivariable Cox proportional hazards regression was used to evaluate time to IV failure. For all survival analyses, IVs with a dwell time within 8 hours before hospital discharge were right censored and this was considered the primary endpoint. Covariates in the multivariable model were determined a priori as known predictors of IV failure or difficult IV access and consisted of end-stage renal disease status, gender, age, and body mass index (BMI). Intravenous drug use (IVDU) status was selected a priori as well but not included in the final analysis due to extremely low prevalence of IVDU in our cohort. Early IV failures, defined as within 4 hours of placement, were excluded from the final multivariable model for two reasons First, these were considered to be a reflection of placement technique rather than catheter durability. Second, removal from the final Cox model significantly improved the proportionality of hazards. Additionally, vein location was not used in our model due to standard of care placement of EDC in the forearm versus multiple potential sites for standard USIV. However, subgroup analyses were conducted limiting analyses to only observations obtained from forearm IVs. In all analyses, each IV was treated as an individual observation ignoring the effects of clustering by patient or encounter. All statistical analyses were performed in RStudio version 1.2.5 (RStudio, Boston, MA) with R version 3.6.2 (The R Foundation, Vienna, Austria).
3. Results
3.1. Participants
Patient and encounter characteristics are detailed in Table 1 and Figure 1. A total of 4,549 ultrasound guided IVs were placed during the study period and 359 of these IVs were EDC. There were no significant differences between groups in age, gender, race, BMI and comorbidities. ESRD patients were more likely to receive an EDC. IVDU was extremely low prevalence in this cohort. Of note, black and female patients were enriched in this cohort relative to the ED population during the study period where black patients represented approximately 17% of encounters and females were approximately 54%.
Table 1.
Patient Characteristics
| USIV | EDC | P-value* | |
|---|---|---|---|
| Peripheral IVs, n (%) | 4190 (92.1%) | 359 (7.9%) | |
| ED Encounters, n (%) | 3668 (91.3%) | 348 (8.7%) | |
| Patients, n (%) | 2669 (89.0%) | 330 (11.0%) | |
| Max PIV per ED encounter, n | 4 | 2 | |
| Max encounters per patient, n | 20 | 3 | |
| Max PIV per patient, n | 24 | 3 | |
| Female, n (% of PIVs) | 2634 (62.9%) | 230 (64.1%) | 0.69 |
| Race | |||
| White, n (% of PIVs) | 2800 (66.8%) | 246 (68.5%) | |
| Black, n (% of PIVs) | 1214 (29.0%) | 95 (26.5%) | |
| Asian, n (% of PIVs) | 59 (1.4%) | 9 (2.5%) | |
| Other, n (% of PIVs) | 94 (2.2%) | 5 (1.4%) | |
| Not Reported, n (% of PIVs) | 23 (0.1%) | 4 (1.1%) | |
| Age, Median years (IQR) | 57.9 (44.2 – 69.5) | 57.7 (41.7 – 71.5) | 0.58 |
| BMI, Median kg/m2 (IQR)# | 28.7 (23.7 – 35.4) | 29.4 (24.0 – 36.9) | 0.35 |
| Charlson Comorbidity Score, Median (IQR) | 4 (1 – 7) | 4 (2 – 7) | |
| ESRD, n (% of PIVs) | 536 (12.8%) | 64 (17.8%) | 0.01 |
| IVDU, n (%of PIVs)$ | 23 (0.5%) | 1 (0.3%) |
P-value is reported only for a priori selected covariates in subsequent analyses
BMI not documented in 1 patient encounter (1 PIV)
IVDU status not reported in 77 patient encounters (89 PIVs)
3.2. IV characteristics
IV characteristics are detailed in Table 2. The majority of IVs in this cohort were placed by ED technicians (98.3% EDC and 95.6% USIV) and on the first attempt (92.8% EDC and 91.8% USIV). There were no significant differences between groups for operator or first pass insertion success rate. Catheter length, size and placement location were different between groups by protocol.
Table 2.
PIV Insertion Characteristics
| USIV | EDC | P-value | |
|---|---|---|---|
| Inserting Provider, n (% of Group) | |||
| ED Technician | 4007 (95.6%) | 353 (98.3%) | 0.82* |
| Physician | 60 (1.4%) | 5 (1.4%) | |
| Nurse | 6 (0.1%) | ||
| Not Documented | 117 (2.8%) | 1 (0.3%) | |
| Insertion Attempts, (% of Group) | |||
| 1 | 3846 (91.8%) | 333 (92.8%) | 0.62# |
| 2 | 224 (5.4%) | 19 (5.3%) | |
| 3 | 25 (0.6%) | 4 (1.1%) | |
| 4 | 3 (0.1%) | ||
| Not Documented | 92 (2.2%) | 3 (0.8%) | |
| Catheter Length, n (% of Group) | |||
| < 1.5 in | 250 (6.0%) | ||
| > 1.5 in | 3687 (88.0%) | ||
| 6 cm | 313 (87.2%) | ||
| 8 cm | 34 (9.5%) | ||
| 10 cm | 3 (1.1%) | ||
| Not Documented | 253 (6.0%) | 9 (1.4%) | |
| Catheter Gauge, n (% of Group) | |||
| 14g | 2 (0.1%) | ||
| 16g | 31 (0.7%) | ||
| 18g | 769 (18.3%) | 11 (3.1%) | |
| 20g | 3300 (78.8%) | 165 (46.0%) | |
| 22g | 64 (1.5%) | 177 (49.3%) | |
| 24g | 1 (0%) | ||
| Not Documented | 26 (0.6%) | 6 (1.7%) | |
| Location, n (% of Group) | |||
| Neck | 2 (0.1%) | ||
| Upper Arm | 1225 (29.3%) | 13 (3.6%) | |
| Antecubital Fossa | 387 (9.2%) | 1 (0.3%) | |
| Brachial | 165 (3.9%) | ||
| Forearm | 2092 (49.9%) | 302 (84.1%) | |
| Wrist | 193 (4.6%) | 31 (8.6%) | |
| Hand | 10 (0.2%) | ||
| Lower Leg | 35 (0.8%) | 7 (2.0%) | |
| Foot or Ankle | 2 (0.1%) | ||
| Other | 9 (0.2%) | 2 (0.6%) | |
| Not Documented | 70 (1.7%) | 3 (0.8%) |
ED technician placed versus other
First pass success versus more than 1 attempt
3.3. Peripheral IV Dwell Time
In univariate survival analyses, median catheter dwell time was 3.8 days (95% CI 3.6 – 4.0) for standard USIV and 5.9 days (95% CI 5.1 – 6.7) for EDC (Table 3 and Figure 2A). A higher percentage (48.2% vs. 42.7%) of EDC survived to discharge and a lower percentage (EDC: 1.1% vs. USIV: 4.2%) failed within 4 hours after placement and a higher percentage of EDCs (9.5% vs. 6.6%) did not have a removal time specifically documented. A multivariable Cox proportional hazards model was used to estimate an adjusted median dwell time of 4.0 days (95% CI 3.8 – 4.1) for standard USIV and 5.9 days (95% CI 5.1 – 6.7) for EDC (Figure 2B). EDC demonstrated significantly improved survival with a HR 0.70 (95% CI 0.60 – 0.81) when adjusted for ESRD, gender, BMI and age (Figure 3). Older age and male gender were also significantly associated with decreased risk of failure. In this model, we removed early failure (dwell time < 4 hours) observations and one observation was missing BMI data which was also subject to listwise deletion. All covariates included the final model did not demonstrate collinearity and the final model did not violate the proportional hazards assumption. Subgroup analysis of IVs placed in the forearm and exclusion of short (< 1.5 inch) standard USIV did not significantly alter these results. These data are presented in the supplemental figures (Figure S2 and S3).
Table 3.
PIV Outcomes
| USIV | EDC | P-value | |
|---|---|---|---|
| Median dwell time, Days (95% CI)* | 3.8 (3.6 – 4.0) | 5.9 (5.1 – 6.7) | < .0001 |
| Median hospital LOS, Days (IQR) | 4.5 (2.4 – 8.1) | 5.2 (3.0 – 9.5) | 0.018 |
| PIV survival to discharge, n (%)# | 1791 (42.7%) | 173 (48.2%) | 0.046 |
| PIV failure within 4 hours, n (%) | 177 (4.2%) | 4 (1.1%) | 0.002 |
| PIV removal time not documented$ | 277 (6.6%) | 34 (9.5%) | 0.039 |
| VAST PIV charges, n (#/encounter) | 3643 (0.99) | 291 (0.84) | < .0001 |
| Encounters with zero VAST charges, n (% of encounters) | 2289 (62.4%) | 227 (65.2%) | 0.298 |
| PICC line insertions, n (% of encounters) | 309 (8.4%) | 28 (8.0%) | 0.808 |
| CT contrast injections, n | 2490 | 214 | |
| CT contrast extravasations, n (% of injections) | 2 (0.08%) | 0 (0%) | 1.00 |
Median dwell time was estimated using the Kaplan-Meier method
Survival to discharge defined as survival to within 8 hours of discharge
For PIVs without documented removal time, hospital discharge time was substituted
Figure 2.
Unadjusted and adjusted Kaplan-Meier survival curves. A) Unadjusted survival analysis curve with all USIV and EDC in the cohort included for analysis. Risk table shown below. B) Adjusted survival curve where early PIV failures (< 4 hours) and covariate means (age 56.7 and BMI 30.8) or reference condition (absence of ESRD and female) are fixed.
Figure 3.
Forest Plot for Hazard of Peripheral IV (PIV) Failure. A multivariable Cox proportional hazards model was constructed using time to failure as the outcome of interest and PIV type, ESRD status, female gender, BMI and age (per 10 years) as predictor variables. Extended dwell catheter (EDC) and higher age predicted longer catheter dwell time. Female gender predicted shorter catheter dwell time.
3.4. Other Peripheral IV Outcomes
There were fewer VAST charges per encounter in the EDC group (0.84 charges/encounter in EDC versus 0.99/encounter in USIV, p < .0001, Table 3). In approximately 45% of total observations, a removal reason was documented (Chi-sq p-value = 0.313). Documented removal reasons (Table 4) were similar between USIV and EDC groups but not analyzed for statistical significance due to limited numbers of observations in the EDC group. First insertion attempt success was also similar between groups. Finally, there were no documented CT contrast extravasations in the EDC group and 2 in the USIV group.
Table 4.
Removal Reason
| USIV | EDC | |
|---|---|---|
| Catheter damage, n (%) | 165 (3.9%) | 13 (3.6%) |
| Change in patient condition, n (%) | 35 (0.8%) | 3 (0.8%) |
| Change in site condition, n (%) | 108 (2.6%) | 9 (2.5%) |
| Damage to catheter, n (%) | 79 (1.9%) | 5 (1.4%) |
| Drainage, n (%) | 69 (1.6%) | 7 (1.9%) |
| Infiltrated, n (%) | 275 (6.6%) | 13 (3.6%) |
| Occluded, n (%) | 292 (7.0%) | 30 (8.4%) |
| Other, n (%) | 270 (6.4%) | 28 (7.8%) |
| Patient discharged, n (%) | 212 (5.1%) | 17 (4.7%) |
| Per order, n (%) | 520 (12.4%) | 51 (14.2%) |
| Per patient/family request, n (%) | 157 (3.7%) | 5 (1.4%) |
| Planned change, n (%) | 20 (0.5%) | 4 (1.1%) |
| Unplanned by other, n (%) | 33 (0.8%) | 1 (0.3%) |
| Unplanned by patient, n (%) | 133 (3.2%) | 7 (1.9%) |
| Not documented, n (%) | 1822 (43.5%) | 166 (46.2%) |
4. Discussion
Our results demonstrate that EDC placements, primarily performed by paramedic trained ED technicians, have superior dwell time to standard USIV even after adjustment for age, gender, BMI and ESRD status. Prior studies have called into question the durability of USIV citing shorter dwell times and increased complications7. Of note, in a small randomized controlled trial, EDC were also shown to have superior dwell times to standard USIV (4.04 vs 1.25 days median survival)8. While our results are similar, key differences between our study and Bahl et al.8 are differences in inserting operator and significantly larger sample size in our study (>4,000 USIV and 359 EDC versus 33 and 37 respectively). Similarly, in this study and in others evaluating USIV to non-ultrasound guided IVs3,10, the dwell times of standard USIV were approximately 2 days suggesting a significant difference between our baseline USIV survival and that of other centers. Several important differences in our study that may explain this discrepancy are inclusion of only patients admitted to the hospital in our study, a long history of the use of USIV in our ED, and maintenance after admission of all peripheral IVs by a dedicated vascular access team. Additionally, our study’s relatively long mean hospital length of stay (approximately 5 days) may have influenced dwell times. The etiology of increased dwell time of EDC is likely multifactorial and due in part to these factors as well as the longer catheter length which has been previously shown to improve catheter dwell time10. In addition to increased dwell time, we have also shown that EDC catheters result in fewer vascular access charges per encounter, without a significant increase in other complications including CT contrast extravasation, and similar differences in removal reason or PICC line insertions while maintaining a high level of first pass insertion success. Interestingly, in our multivariable model for time to IV failure, ESRD status exhibited a trend towards reduced hazard of failure which reached statistical significance when examining only forearm placed IVs (Figures 3 and S2). The etiology of this effect is unclear although likely due to increased efforts by all patient care teams to preserve IV access and longer lengths of stay in ESRD patients. Age also exhibited this protective effect in our model which we believe is due to similar reasons. Overall, our results demonstrate improved durability of EDC compared to standard USIV even in our setting where baseline dwell time is high and complication rates are relatively low. Given increasing popularity of midline catheter use both in admitted and ED patients11,12, as well as increasing emphasis on appropriate selection of vascular access to match duration and type of infusion therapy9, EDC may represent a new alternative in the spectrum of vascular access devices.
4.1. Limitations
Our study has several limitations. First, measurement of dwell time was calculated as days between documented placement and removal. However, in approximately 9.5% of EDC and 6.6% of standard USIV no removal time was documented. These observations continued to have other IV maintenance or medication charting and were treated as if they had survival to hospital discharge. The survival analysis with these observations removed did not result in significant changes (Figure S4). Secondly, while the vast majority of EDC and USIV are placed by ED technicians in our DIVA team, a handful were placed by physicians or nurses. These observations represent a tiny fraction of total IVs in this cohort and thus were not expected to modify results. Prior studies have also demonstrated that placement location can adversely affect dwell times of ED placed peripheral IVs13. By departmental policy, EDC are placed almost exclusively in the forearm while USIV have more heterogeneity in site selection (Table 3). Similarly, there is more heterogeneity in catheter size of USIV when compared to EDC which are predominantly 20g or 22g. DIVA team members use their discretion based on clinical judgement when selecting EDC or standard USIV which likely introduces selection bias based on the perceived need for longer term peripheral IV access or difficulty of IV insertion. This may be the reason for enrichment of ESRD patients in the EDC group and longer observed hospital length of stay. Prior studies have demonstrated that insertion depth is a significant predictor for IV failure and, in our institution insertion depth from skin to target vein is not documented. Thus, this variable could not be analyzed. However, the postulated mechanism for failure of IVs placed in deeper veins is due to short catheter length actually inserted into the vein which is less likely in our department given policy limiting insertion depth to no more than one third the catheter length and use of predominantly long (4.5 cm) IV catheters in the standard USIV group. Finally, in our multivariable analysis, early IV failures (dwell time < 4 hours) were removed because these failures were considered to be failed insertion technique rather than reflective of catheter durability and their removal from analysis improved proportionality of hazards in the Cox model. A higher percentage of early failures occurred in the USIV group compared to EDC (4.4% versus 1.1%, Table 3) and our overall results were not altered by retention of these observations in the multivariable analysis.
4.2. Generalizability
There are also several factors that limit the generalizability of our findings. First, in our institution all ED placed IVs are maintained by a dedicated vascular access team after admission. This is unlike other institutions where bedside nurses or medical assistants may be responsible for IV maintenance. Also, by institutional policy, blood draws from existing peripheral IVs are prohibited after patients leave the ED. These two factors may contribute to longer dwell times seen at our institution relative to others. Our institution is also a tertiary care center with a high prevalence of difficult access including ESRD patients which may limit applicability to other centers where difficult IV access is less common. As a result of this, our department has employed a paramedic trained, ED technician run DIVA team since 2008 and all current members of this team are highly experienced. At other centers where physicians or nurses primarily perform ultrasound guided IV access these results may not be applicable. Finally, the retrospective nature of this study and reliance on EMR flowsheet documentation may be inherent limitations on generalizability.
5. Conclusions
Extended dwell peripheral IVs are a superior alternative to standard long IV catheters with improved dwell time and a very low rate of complications when placed by emergency department technicians.
Supplementary Material
Figure S1. Endurance Extended Dwell Peripheral Catheter System. Teleflex, Morrisville, NC
Figure S2. Forest Plot for Hazard of Peripheral IV (PIV) Failure in Forearm PIV. A similar multivariable model was used to estimate hazard of failure in subgroup of observations consisting of only forearm placed PIV. 2032 observations (57 EDC and 1975 standard USIV) were excluded relative to the final model in the primary analysis. The overall effects and relative effect sizes are unchanged although the effect of ESRD status has become statistically significant.
Figure S3. Forest Plot for Hazard of Peripheral IV (PIV) Failure with Short PIVs Excluded. A similar multivariable model was used to estimate hazard of failure in subgroup of observations with short (< 1.5 in) IVs excluded. 231 observations (all standard USIV) were excluded relative to the final model in the primary analysis. The overall effects and relative effect sizes are unchanged.
Figure S4. Forest Plot for Hazard of Peripheral IV (PIV) Failure with Missing Removal Times Excluded. A similar multivariable model was used to estimate hazard of failure in subgroup of observations consisting of only those with documented removal times. 299 observations (34 EDC and 265 standard USIV) were excluded relative to the final model in the primary analysis. The overall effects and relative effect sizes are unchanged.
Acknowledgements:
The authors acknowledge and thank the ED technicians, nurses, physicians and physician assistants of the Department of Emergency Medicine at Michigan Medicine. The authors also acknowledge and thank the Vascular Access Services Team at Michigan Medicine.
Funding Sources/Disclosures:
CF has received research support unrelated to this work from the National Heart, Lung, and Blood Institute (1K12HL133304). NT has received research support unrelated to this work from GE Healthcare Inc.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Prior presentations: Components of this work have previously been presented at the SAEM 2020 Annual Meeting.
Conflicts of Interest Disclosure: JT reports that has received payment from Teleflex Inc. as a clinical educator.
References
- [1].Davis EM, Feinsmith S, Amick AE, Sell J, McDonald V, Trinquero P, et al. Difficult intravenous access in the emergency department: Performance and impact of ultrasound-guided IV insertion performed by nurses. Am J Emerg Med. 2020;0(0). doi: 10.1016/j.ajem.2020.11.013 [DOI] [PubMed] [Google Scholar]
- [2].Au AK, Rotte MJ, Grzybowski RJ, Ku BS, Fields JM. Decrease in central venous catheter placement due to use of ultrasound guidance for peripheral intravenous catheters. Am J Emerg Med. 2012;30(9):1950–1954. doi: 10.1016/J.AJEM.2012.04.016 [DOI] [PubMed] [Google Scholar]
- [3].Adhikari S, Blaivas M, Morrison D, Lander L. Comparison of infection rates among ultrasound-guided versus traditionally placed peripheral intravenous lines. J Ultrasound Med. 2010;29(5):741–747. doi: 10.7863/jum.2010.29.5.741 [DOI] [PubMed] [Google Scholar]
- [4].Dargin JM, Rebholz CM, Lowenstein RA, Mitchell PM, Feldman JA. Ultrasonography-guided peripheral intravenous catheter survival in ED patients with difficult access. Am J Emerg Med. 2010;28(1):1–7. doi: 10.1016/j.ajem.2008.09.001 [DOI] [PubMed] [Google Scholar]
- [5].Rupp JD, Ferre RM, Boyd JS, Dearing E, McNaughton CD, Liu D, et al. Extravasation Risk Using Ultrasound-guided Peripheral Intravenous Catheters for Computed Tomography Contrast Administration. Jang TB, ed. Acad Emerg Med. 2016;23(8):918–921. doi: 10.1111/acem.13000 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [6].Favot M, Gallien J, Malik A, Kasten A, Wells R, Ehrman R. Contrast Extravasation as a Complication of Emergency Nurse-Performed Ultrasound-Guided Peripheral Intravenous Catheter Placement. J Emerg Nurs. 2019;45(5):512–516. doi: 10.1016/j.jen.2019.05.016 [DOI] [PubMed] [Google Scholar]
- [7].Witting MD, Moayedi S, Dunning K, Babin LS, Cogan BM. Power Injection Through Ultrasound-Guided Intravenous Lines: Safety and Efficacy Under an Institutional Protocol. In: Journal of Emergency Medicine. Vol 52. Elsevier; USA; 2017:16–22. doi: 10.1016/j.jemermed.2016.09.017 [DOI] [PubMed] [Google Scholar]
- [8].Bahl A, Hang B, Brackney A, Joseph S, Karabon P, Mohammad A, et al. Standard long IV catheters versus extended dwell catheters: A randomized comparison of ultrasound-guided catheter survival. Am J Emerg Med. 2019;37(4):715–721. doi: 10.1016/j.ajem.2018.07.031 [DOI] [PubMed] [Google Scholar]
- [9].Chopra V, Flanders SA, Saint S, Woller SC, O’Grady NP, Safdar N, et al. The Michigan Appropriateness Guide for Intravenous Catheters (MAGIC): Results From a Multispecialty Panel Using the RAND/UCLA Appropriateness Method. Ann Intern Med. 2015;163(6_Supplement):S1. doi: 10.7326/M15-0744 [DOI] [PubMed] [Google Scholar]
- [10].Pandurangadu AV, Tucker J, Brackney AR, Bahl A. Ultrasound-guided intravenous catheter survival impacted by amount of catheter residing in the vein. Emerg Med J. 2018;35(9):550–555. doi: 10.1136/emermed-2017-206803 [DOI] [PubMed] [Google Scholar]
- [11].Chopra V, Kaatz S, Swaminathan L, Boldenow T, Snyder A, Burris R, et al. Variation in use and outcomes related to midline catheters: Results from a multicentre pilot study. BMJ Qual Saf. 2019;28(9):714–720. doi: 10.1136/bmjqs-2018-008554 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [12].Spiegel RJ, Eraso D, Leibner E, Thode H, Morley EJ, Weingart S. The Utility of Midline Intravenous Catheters in Critically Ill Emergency Department Patients. Ann Emerg Med. 2020;75(4):538–545. doi: 10.1016/j.annemergmed.2019.09.018 [DOI] [PubMed] [Google Scholar]
- [13].Fields JM, Dean AJ, Todman RW, Au AK, Anderson KL, Ku BS, et al. The effect of vessel depth, diameter, and location on ultrasound-guided peripheral intravenous catheter longevity. Am J Emerg Med. 2012;30(7):1134–1140. doi: 10.1016/j.ajem.2011.07.027 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Figure S1. Endurance Extended Dwell Peripheral Catheter System. Teleflex, Morrisville, NC
Figure S2. Forest Plot for Hazard of Peripheral IV (PIV) Failure in Forearm PIV. A similar multivariable model was used to estimate hazard of failure in subgroup of observations consisting of only forearm placed PIV. 2032 observations (57 EDC and 1975 standard USIV) were excluded relative to the final model in the primary analysis. The overall effects and relative effect sizes are unchanged although the effect of ESRD status has become statistically significant.
Figure S3. Forest Plot for Hazard of Peripheral IV (PIV) Failure with Short PIVs Excluded. A similar multivariable model was used to estimate hazard of failure in subgroup of observations with short (< 1.5 in) IVs excluded. 231 observations (all standard USIV) were excluded relative to the final model in the primary analysis. The overall effects and relative effect sizes are unchanged.
Figure S4. Forest Plot for Hazard of Peripheral IV (PIV) Failure with Missing Removal Times Excluded. A similar multivariable model was used to estimate hazard of failure in subgroup of observations consisting of only those with documented removal times. 299 observations (34 EDC and 265 standard USIV) were excluded relative to the final model in the primary analysis. The overall effects and relative effect sizes are unchanged.